116 M. Garcia et al. Virus Research 51 (1997) 115 124
raminidase (N1-N9) subtypes have been isolated from migratory waterfowl in North America as indicated by surveillance studies from 1974 to 1988 (Slemons et al., 1974; Slemons and Easter- day, 1977; Hinshaw et al., 1980; Halvorson et al.~ 1982; Deibel et al., 1985; Nettles et al., 1985; Stallknecht et al., 1990; Slemons et al., 1991: Graves, 1992). Initially, migratory waterfowl were proposed to play a major role in the transport and dissemination of influenza viruses (Slemons et al., 1974). The widely distributed influenza virus pool observed in migratory waterfowl later sug- gested these as the host reservoir for influenza viruses that cause outbreaks in domestic poultry (Hinshaw et al., 1980). These proposals are con- sistent with the common occurrence of low patho- genicity (LP) outbreaks in poultry. For example, LP outbreaks caused by influenza viruses of sev- eral subtypes are common among range-reared turkey flocks in Minnesota, where the outbreaks coincide with an increase in waterfowl popula- tions during their southward migration through the state (Halvorson et al., 1982). In contrast to the wide variety of AI subtypes involved in out- breaks of low or moderate pathogenicity, highly pathogenic (HP) AI outbreaks have been re- stricted to a few viruses of the H5 and H7 sub- types (Wood et al., 1993; Senne et al., 1996a).
Although the expression of the HP phenotype among AI viruses has been linked to multiple genes (Rott et al., 1979), structural features in the hemagglutinin glycoprotein are a primary deter- minant of an isolate's pathogenicity (Webster and Rott, 1987; Kawaoka and Webster, 1988). Analy- ses of H5 viruses isolated during outbreaks in poultry in North America have demonstrated that HP viruses most likely emerge from LP forms of the virus after acquiring specific mutations in the HA gene (Kawaoka et al., 1984; Horimoto et al., 1995a; Garcia et al., 1996). Previous phylogenetic analysis of the HA gene, based upon partial HA] sequences from H5 viruses, provided additional evidence indicating that HP H5 viruses do not constitute unique lineages but rather arise from LP precursors introduced into poultry probably from aquatic birds (Rohm et al., 1995).
Three outbreaks of highly pathogenic AI of the H5 subtype have occurred in North America. The
prototype H5 viruses were isolated from turkeys in Ontario during 1966 where the virus caused a limited HP outbreak (Lang et al., 1968). The second HP outbreak was in Pennsylvania during 1983 and 1984 (Eckroade and Bachin, 1987), while the third occurred more recently in central Mexico during 1994 to 1995 (Senne et al., 1996b). Precise identification of sources of virus causing HP outbreaks in poultry has remained elusive. The sporadic isolation of H5N2 viruses from chickens in live poultry markets since 1986 sug- gests that in addition to turkeys, these birds may be a potential source of H5 LP viruses for poultry (Senne et al., 1992). In order to identify all the potential reservoirs of H5 influenza viruses infect- ing poultry it is important to study the genetic relationship among viruses from all available hosts including aquatic birds, live market birds and commercial poultry. This will allow a clearer understanding of the epidemiological significance of each host in the generation of HP AI out- breaks. This study presents analysis of the HAj and the NS gene sequences of H5 viruses isolated in North America during the past 29 years. It demonstrates that avian influenza viruses of the H5 subtype continue to undergo dynamic evolu- tionary variation including apparent variation in mutation rates among different host species and gene re-assortment among mixed virus popula- tions in the same host species.
2. Materials and methods
2.1. Viruses
A total of 26 H5 viruses were included in these phylogenetic analysis: 22 from North America and four isolates from Europe. Among the North America viruses, six were from migratory water- fowl, one isolate from shore birds, one from emu, two from live poultry market chickens, four from turkeys and eight from commercial chickens. In this study sequences of the HA1 plus the N-termi- nal six amino acids of the fusion peptide were obtained for six waterfowl, three turkey, and one chicken isolate. The complete sequence for the NS gene RNA was obtained for 20 isolates. The
M. Garcia et al . /Virus Research 51 (1997) 115-124
Table 1 H5 avian influenza virus isolates utilized in the phylogenetic analysis
117
Viruses Highly Abbreviation HA sequence accession NS sequence accession pathogenic number number
A/turkey/Ontario/7732/66(H5N9) Yes A/turkey/Wisconsin/68(H 5N9) No A/mallard/Wisconsin/34/75(H5N6) No A/mallard/Wisconsin/169/75(H5N3) No A/mallard/Wisconsin/428/75(H 5N 3) No A/duck/Michigan/80(H5N2) No A/turkey/Minnesota/3689-1551 / No
81(H5N2) A/chicken/Pennsylvania/1/83(H5N2) No A/chicken/Pennsylvania/1370/ Yes
83(H5N2) A/mallard/Ohio/556/87(H5N9) No A/mallard/Ohio/345/88(H5N2) No A/ruddy turnstone/Delaware/244/ No
91(H5N2) A/chicken/Pennsylvania/13609/ No
93(H5N2) A/chicken/Florida/25717/93(H5N2) No A/emu/Texas/399243/93(H5N2) No A/chicken/Mexico/31381-7/94(H5N2) No A/chicken/Hidalgo/26654-1368/ No
94(H5N2) A/chicken/Mexico/31381-1/94(H5N2) No A/chicken/Mexico/3138 I-2/94(H5N2) No A/chicken/Queretaro/14588-19/ Yes
TOn66 ~ M30122 h U85376 g TWi68 a U79456 g U85378 g MWi75A b U79451 g U85379 g MWi75B b U79452 g U85375 g MWi75C b U79453 g U85380 g DMi80 ~ U79449 g U853811 g TMn81 a U79454 g U85382 g
CPe83E a M 18001h NA i CPe83L a M10243 h NA i
MOh87 ¢ U67783 g U85377 g MOh88 c U79450 g NA i RDe91 UO5330 h NA i
CPe93 ~ U05331 h U85383 g
CF193 U05332 h NA i ETx93 a U28919 h U85384 g CMx93 ~ U37165 h U85385 g CHi94 ¢ U37172 h U85386 g
CJa94A ~ U37166 h U85387 g CJa94B ¢ U37167 h U85388 g CQ95A ~ U37182 h U85389 g
CQ95B ¢ U79448 g U85390 g
TMn95 ~ U79455 g U85391 g CS59 X07826 h AF009898 g D183 M 18450 h NA i TI83 f Ml8451h U85392 g TE91 f Wood et al., 1994 U85447 g
a Viruses obtained from the repository of the Southeast Poultry Research Laboratory, USDA, ARS (Athens, GA). b Viruses isolated by Slemons and Easterday; obtained from the repository of Dr. B.C. Easterday at University of Wisconsin (Madison, WI). c Viruses obtained from the repository of Dr. R. Slemons at Ohio State University (Columbus, OH). d Viruses obtained from the repository of Dr. R. Webster at St. Jude Children's Research Hospital (Memphis, TN).
Viruses obtained from the repository of the National Veterinary Service Laboratory (NVSL) (Ames, IA). fViruses obtained from the repository of Dr. D. Alexander at Central Veterinary Laboratory (Surrey, UK). g Sequences obtained during this study. h Sequences obtained from GeneBank. i Not available.
r e m a i n i n g s e q u e n c e s w e r e o b t a i n e d f r o m G e n -
B a n k . A c o m p l e t e l i s t o f t h e v i r u s e s a n d s e q u e n c e s
u t i l i z e d in t h i s s t u d y a r e p r e s e n t e d in T a b l e 1.
V i r u s e s w e r e p r o p a g a t e d in t h e a l l a n t o i c c a v i t y o f
1 0 - d a y - o l d c h i c k e n e m b r y o s a s p r e v i o u s l y d e -
s c r i b e d b y P e r d u e e t al . 0 9 9 0 ) .
118 M. Garcia et a l . / Virus Research 51 (1997) 115-124
2.2. Sequence analysis
Viral RNA extraction, amplification by reverse transcriptase-polymerase chain reaction (RT- PCR) and sequencing of PCR products of the HA gene were performed as previously described by Garcia et al., 1996. The NS gene of selected isolates was amplified by RT-PCR using primers consisting of the conserved 5' terminal 12 bases and the 3' terminal 13 bases coupled to a 12 base extension containing four uracils. The amplified PCR product containing the full length NS gene cDNA was extracted from an agarose gel and inserted into a plasmid vector, (pAMP1; Life Technologies, Gaithersburg, MD). The plasmid was used to transform DH5~ cells (Life Technolo- gies) and positive colonies were screened with cDNAs specific for the NS gene. Positive colonies were grown overnight and plasmid extracted with the High Pure Plasmid Isolation Kit (Boehringer Mannheim, Indianapolis, IN). All sequences were obtained in both directions on an ABI automated sequencer (Smith et al., 1986) using the manufac- turer's suggested protocols. Primer sequences are available upon request.
2.3. Sequence data analysis
Nucleotide sequences, prediction of amino acid sequences and alignments were completed using the GeneWorks 2.45 software (Intelligenetics, Mountain View, CA), or DNASTAR (Madison, WS). Phylogenetic analysis and pair-wise dis- tances between gene sequences were calculated with PAUP software, Version 3.1.1 (Swofford, 1989). Phylogenetic trees were generated utilizing heuristic searches with the neighbor-joining method. Error associated with tree structures was evaluated by bootstrap analyses.
3. Results and discussion
Analysis of phylogenetic relationships among the HA genes based on nucleotide and amino acid sequences exhibited a divergence of H5 subtypes into the two geographically distinct lineages previ- ously reported (Rohm et al., 1995), a North American and a European lineage. Although the
topology and distribution of sequences in phylo- genetic trees (phylograms) based on nucleotides and amino acids varied slightly, four sub-lineages or clades were easily defined for the North Amer- ican viruses. The groups are designated as I-IV (Fig. 1). Clade I includes the prototype highly pathogenic H5 isolate from Ontario in 1966, a turkey isolate, a waterfowl origin virus MWi75A and the two 1983 Pennsylvania outbreak viruses (CPe83E, CPe83L) which share a common ances- tor. Clade II includes primarily duck isolates with sequences (MWi75B and MWi75C) that clearly share a common ancestor with the MOh88 se- quence. The DMi80 and TMn81 isolates included in this clade share a common ancestor as illus- trated by the clustering in the nucleotide con- structed phylogram, but in the amino acid phylogram TMn81 was less distinctly classified. As such, the synonymous changes indicate a sin- gle clade where the non-synonymous changes in- dicate separate clades. Clade III has sequences from shore bird (RDe91), live poultry market chickens (CF193, CPe93), ratite (ETx93) and turkey (TMn95) viruses sharing a common ances- tor. Clade IV includes sequences from viruses isolated during the recent outbreak of HP AI in Mexico. Within this sub-lineage viruses CQ95A and CQ95B were characterized as highly patho- genic. In the interest of streamlining data presen- tation, only a few Mexican isolates were selected to represent the diversity of the 18 viruses previ- ously characterized in this clade (Garcia et al., 1996). In both phylograms, clade IV sequences from HP and LP viruses share an apparent com- mon ancestor. Waterfowl and shore bird virus sequences are found among three of the four sub-lineages indicating that multiple H5 virus sub- lineages are maintained in aquatic birds. Further- more, sequences from waterfowl viruses MWi75A, MWi75B and MWi75C, isolated during the same year by the same laboratory, are found in two clearly separated clades of the phylogenetic tree.
The results illustrated in Fig. 1 clearly show that multiple populations or lineages of H5 viruses are present at the same time in the same geographic location in a waterfowl population. Pair-wise distances between H5 subtype influenza viruses from migratory waterfowl and poultry
M. Garcia et a l . / Virus Research 5l (1997) 115-124 119
A
North America
TOn66 _ ~ T W i 6 8
MWi75A CPe83E CPe83L
._~MWi75B -~ MWi75C
MOb88 DMi80 II
"~" TMn81
-MOh87 1 RDe91 [ - - CPe93
CF93 L . . ET93 -- TMn95
Ill
B
~ C CMx93 CJa94A CJa94B
CQS,SA IV CQgSB Hi94
CS59 DI83 "1"183 Europe ~' TE91 Europe
North America
TOn66 [~ ["- MWlTSA
168 "[ CPe83L ~ - ~ "rw CPe83E
I .1" MWI75B I . . , 5 c [ ~l-.. MOh88
11' ° ' " I I t TMn81 IJI r cP°93 I I~"~] CFL9a
- r TMng5 It._ RDe91 • MOh87
~cC~Q Mx93 a94B CHi94 94 95A
CQ95B CS59
. ~ T E 9 1 DI83 TI83
Fig. 1. Phylogenetic trees of the HA gene from H5 influenza viruses. (A) The phylogram constructed for 26 H5 viruses utilizing complete nucleotide coding sequences of the HAl subunit and the N-terminal fusion peptide sequences. The tree was rooted to the H2 HA I sequence of A/mallard/Montana/61 (H2N2) (GeneBank accession number L11136). A single nucleotide tree was obtained after 2000 bootstrap re-samplings. (B) The amino acid consensus tree from 15 trees obtained after 1000 bootstrap re-samplings. The lengths of the horizontal lines are proportional to the nucleotide and amino acid changes between sequences. Vertical lines separate progeny virus lineages at the point where they branch from a theoretical common ancestor. The arrow at the left indicates the direction of the A/mallard/Montana/61 (H2N2) from the root node. Abbreviations of viruses are listed in Table 1.
origin viruses indicate that a close phylogenetic relationship exists, particularly between sequences from turkey and waterfowl origin viruses. Nucle- otide distance comparisons between HA1 se- quences of waterfowl and turkey origin viruses showed that viruses MOh87 and TMn95 have the highest HAl nucleotide sequence similarity, with a nucleotide pair-wise distance of 1.7%. However, MOh87 and TMn81 have the highest HA1 amino acid similarity, followed by MOh87 and TMn95. The amino acid pair-wise distance shared by MOb87 and TMn81 is 1.8%, and 2.1% for MOh87 and TMn95. Temporal and spatial epi- demiological evidence has linked influenza infec- tions of turkeys in Minnesota with viruses isolated from waterfowl in the area (Halvorson et al.,
1982). The high sequence similarity observed be- tween the waterfowl origin virus MOh87 and the Minnesota turkey viruses (TMn81 and TMn95) HAl sequences is the first molecular epidemiologi- cal evidence that links influenza H5 infection in Minnesota turkeys with a waterfowl origin virus. Furthermore, the high level of sequence similarity indicates that a stable H5 gene lineage has been maintained in the waterfowl, and that viruses carrying that particular HA gene lineage have been sporadically introduced into Minnesota turkey flocks.
Influenza surveillance studies of live poultry markets have demonstrated that, recently, the most common influenza virus isolated from this environment is the H5N2 subtype, isolated from
120 M. Garcia et al. / Virus Research 51 (1997) 115 124
chickens, ducks, guinea fowl, turkey, quail, pi- geon and chuckar partridge (Senne et al., 1992). This virus subtype reappeared in chickens from live poultry markets in 1993 (Saito et al., 1994). Comparison of nucleotide and amino acid dis- tances between HA~ sequences of waterfowl and all chicken origin viruses demonstrates that MOh87 and CFI93 share the highest HA similar- ity followed by MOh87 and CPe93. The nucle- otide and amino acid pair-wise distances for MOb87 and CF193 were 2.8% and 3.3%, while for MOb87 and CF193 were 2.9% and 3.6%, respec- tively. The close phylogenetic relationship be- tween the HA sequences of waterfowl virus MOb87 and live poultry market chicken viruses, CF193 and CPe93, indicates that a waterfowl virus may have been the source of the 1993 live poultry market viruses.
The apparent differences in the genetic stability, over time, between clade IV and the other clades, led us to consider the mutation rate among these H5 strains. Measuring the mutation rate of avian influenza virus genes has classically involved di- rect comparison of isolates, oftentimes from vari- ous orders of birds, and expressing the variation over time (see Gorman et al., 1992). This may not be a valid approach, however, for measuring an accurate rate as multiple species are involved and multiple introductions of virus have occurred. We previously published a mutation rate for the iso- lates from Mexico based upon the assumption that there was a single introduction and the same virus was allowed to replicate unabated in the poultry populations in Mexico (Garcia et al., 1996). When the clades in Fig. I A were compared, the sub-lineage with the lowest rate of nucleotide and coding changes per year was sub-lineage II, composed of mostly waterfowl viruses isolated during a 13 year period. Clades I and III included sequences from viruses originating from a wider variety of hosts showed slightly higher mutation rate. Though perhaps not accurate, the estimated average mutation rates for H5 virus sub-lineages I, II, and III fall within the estimated evolutionary rates values obtained for human influenza H1 and H3 genes which range from 0.61 to 7.0 x 10 ̀3 nucleotide substitutions per site per year (Gorman et al., 1992). In contrast, clade IV, composed of
sequences from chicken-origin H5N2 viruses from Mexico, acquired 28.1 x 10 .3 nucleotide and 8.8 × 10 ̀3 coding changes per site per year.
These results suggest that nucleotide substitu- tion rates in the HA gene of H5 subtype influenza viruses increase significantly once the virus is in- troduced into commercial poultry. There were only two isolates sequenced from the 1983 poultry outbreak in the northeast US, so it is not known how rapidly these viruses were evolving in the field, The relationships in clade I, however, sug- gest that the CPe83E and CPe83L isolates from this period had undergone significant sequence variation from their progenitor population. At this time, there are no other studies where HA sequences from multiple viruses isolated during a single AI poultry outbreak have been compared. Therefore we cannot presently determine if the higher apparent mutation rate observed for the viruses of Mexican origin is a unique characteris- tic of this particular outbreak or common to any situation in which unrestricted replication in com- mercial poultry is allowed.
The phylogenetic relationship of several sub- type H5 North American isolates based upon the NS gene sequences is shown in Fig. 2. The phylo- gram readily identifies the two previously de- scribed alleles of the NS gene (Treanor et al., 1989; Ludwig et al,, 1991). There also appear to be multiple sub-lineages of viruses circulating in migratory waterfowl based upon the NS sequence data. As in the case of the HA gene data, this is most obvious from the finding that viruses iso- lated at the same time in the same laboratory from a single survey of mallards, yields two dis- tinct alleles of the NS gene. They are found in the isolates which were members of the two separate clades defined by the HA gene phylogram, namely the MWi75A and MWi75C. The fact that the MWi75B isolate contains a hemagglutinin virtu- ally identical to the MWi75C and an NS coding sequence most closely related to the MWi75A A allele is most easily explained by concluding that re-assortment of genes between two distinct sub- lineages of H5 viruses occurred in these wild duck populations. The isolates from Mexico shown in the Fig. 2, as was the case for the HA gene, also demonstrated a higher substitution rate in the NS
M. Garcia et a l . / I/irus Research 51 (1997) 115-124 121
A B TOn66
t r Mwi75A
-4 I ~ TMn95
ETx93 CS59
TI83 Twi68
~ C Mwi75C TMn81 CJa94A CJa94B CHi94 CQ95B CQ95A
DMI80 TE91
r ' - I CJa94A I ~ 1 ' CJa94B
t.- CHi94
I CMx93
CPe93 DMi80
- MWi75C TMn81 TVVi68 TE91
-- MOh87 MWi75B MWi75A TMn95
TOn66 CS59
ETx93 TI83
Fig. 2. Phylogenetic comparison of NS segment cDNA sequences. Several of the isolates used in the HA gene analysis were subjected to cloning and sequencing of the NS RNA. The sequences were analyzed as in Fig. 1. (a) Relationship based on nucleotide sequences. (b) Relationship based on NS1 amino acid sequences. A and B on the cladogram denote the two different alleles previously identified (Treanor et al., 1989).
sequences (19.5 x 10 -3 per site per year) when compared to the published rates of change for the NS gene (1.94 × 10 -3 per site per year and 1.78 × 10 -3 per site per year; see Gorman et al., 1992).
Examination of the HA sequences in this study led to two additional observations regarding the structure of the HA1 segment in H5 subtype viruses. With the exception of the sequences for the TOn66 and the RDe91 isolates, all sequences examined contained completely conserved sequen- ces in the areas surrounding the proposed recep- tor binding site (Fig. 3a). Although it is well known that receptor site sequences are highly con- served among the type A influenza viruses, this is the first report of enough sequences to establish this exclusively in the H5 subtype. The single am- ino acid differences in the TOn66 and RDe91 may be due to observed variation in tissue tropism of these two strains (Van Campen et al., 1989; Saito et al., 1994). The second observation was that all commercial chicken isolates possessed a potential
glycosylation site specified at position 236 which was absent from the other isolates (Fig. 3b).
The number and distribution of glycosylation sites along the HA1 subunit has been implicated in establishing tissue tropism (Inkster et al., 1993), and virulence has been associated with the loss of a glycosylation site at position 11 in field isolates of H5 subtypes (Kawaoka et al., 1984) and in laboratory derivatives passed in 14-day-old chicken embryos (Horimoto et al., 1995b). In this survey, both virulent and avirulent strains were shown to lack the site at asparagine 11, and four potential glysosylation patterns were identified for natural isolates. An alpha-carbon tracing model of the consensus H5 hemagglutinin sequence pre- dicted the position of the potential glycosylation at site 236 was on the surface of the HA molecule globular head near the receptor binding pocket at 220-224 (data not shown). Thus the acquisition of glycosylation at site 236 may be an adaptive fea-
122 M. Garcia et al. / Virus Research 5I (1997) 115 124
a
129 -133 2 2 0 - 2 2 4
b
A
1011 23 168 236 268
10 23 168 236 268 PCS
B
1011 23 168 268
C
10 23 168 268
D I I I ill,, I H I II Fig. 3. Consensus structural features of H5 hemagglutinin genes. (A) Schematic diagram of the HA~ subunit and fusion peptide indicating positions of amino acids predicted to form the receptor binding pocket of the HA molecule for avian influenza H5 virus. Amino acid positions are indicated by the numbers below and above the diagram. Position 1 was given to the first residue of the coding sequence of the HA~ subunit. The proteolytic cleavage site (PCS) is indicated at amino acid position 326. The receptor binding site is formed by amino acids 129-133 (Ser-Gly-Val-Ser-Ser), 220-224 (Asn-Gly-Gln-Ser-Gly), 91 (Tyr), 149 (Trp), 186 (Glu) and 190 (Leu). The dark boxes indicate the conserved amino acids which surround the receptor binding amino acids. Amino acid substitutions were observed at position 129 (Ser-~ Thr) for the HA~ sequence of RDe91 and at position 223 (Ser ~ Asn) for the HA1 sequence of TOn66, The position of the amino acid substitutions are marked with an asterisk. (B) Potential glycosylation site distribution patterns among North American H5 viruses. PCS, proteolytic cleavage site. The amino acid positions for the asparagine (Asn) residue in the potential glycosylation sequences are shown as follows: (A) sequences of CPe83E, CMx93, CJa94A, CJa94B, CQ95A, and CQ95B; (B) sequences of CPe83L and CHi94; (C) sequences of TOn66, TWi68, MWi75A, MWi75B, MWi75C, DMi80, TMn81, MOh87, MOb88, RDe91, CPe93, CF193 and TMn95; (D) ETx93 sequence.
ture of H5 viruses associated with efficient replica- t ion and /or transmissibil i ty in commercial chick- ens. Studies are in progress to determine whether this is the case, as such a marker site might be useful in predict ing the length of time an AI virus has been in a part icular popula t ion of hosts.
These data clearly indicate both a substant ia l level of sequence var ia t ion in the hemagglut in in gene, and re-assortment of gene segments result-
ing in stable sub-lineages in wild waterfowl popu- lations. There is an apparent significant increase in rates of change and possible adaptive changes occurr ing when the viruses enter commercial poultry. It is not know n whether this is due to increased immune pressure, t ransfer between avian orders, a real increase in mu ta t i on rate, or simply due to the increased concent ra t ion of available hosts and increased n u m b e r of isolates.
M. Garcla et al . /Virus Research 51 (1997) 115-'124 123
A n i m p o r t a n t mean ing o f the findings, however , is tha t recent ly p r o p o s e d vacc ina t ion p ro toco l s for cont ro l l ing the H P H5 subtype s trains m a y ult i- ma te ly face the same inherent ant igenic dr i f t p rob l ems faced when vacc ina t ing agains t h u m a n type A influenza viruses.
I t has been prev ious ly suggested ( G o r m a n et al., 1992; Webs t e r et al., 1992) tha t av ian influ- enza viruses are in ' evo lu t iona ry stasis ' , having a d a p t e d to a p r imord i a l reservoir o f feral water - fowl and thus undergo ing negat ive genetic selec- t ion to remain in such an adap ta t ion . These conclus ions were reached based u p o n phyloge- netic analysis o f p r imar i ly the nuc leopro te in (NP) gene (Webs te r et al., 1992), ind ica t ing a lower level o f posi t ive selection in the N P gene se- quences when av ian s trains are c o m p a r e d to hu- m a n or swine strains. However , the d a t a in this r epor t based on the H A and N S genes do no t a p p e a r consis tent wi th the idea o f ' evo lu t iona ry stasis ' . Converse ly , they suggest tha t av ian influ- enza viruses cont inue to evolve at a significant ra te in var ious orders , families, genus and species o f the class Aves.
References
Deibel, R., Emord, D.E., Dukelow, W., Hinshaw, V.S., Wood, M.W., 1985. Influenza viruses and paramyxoviruses in ducks in the Atlantic flyway, 1977-1983, including H5N2 isolate related to the virulent chicken virus. Avian Dis. 29, 970-985.
Eckroade, R.J., Bachin, L.A., 1987. Avian influenza in Penn- sylvania-the beginning. In: Proceedings 2nd International Symposium on Avian Influenza. US Animal Health Asso- ciation, Athens, GA, pp. 22-32.
Garcia, M., Crawford, J., Latimer, J.W., Rivera-Cruz, E., Perdue, M.L., 1996. Heterogeneity in the haemaglutinin gene and emergence of the highly pathogenic phenotype among recent H5N2 avian influenza viruses in Mexico. J. Gen. Virol. 77, 1493-1504.
Graves, I.V., 1992. Influenza viruses in birds of the Atlantic flyway. Avian Dis. 36, 1-10.
Halvorson, D., Karunakaran, D., Senne, D., Kelleher, C., Bailey, C., Abraham, A., Hinshaw, V.S., Newman, J., 1982. Epizootiology of avian influenza-simultaneous moni- toring of sentinel ducks and turkeys in Minnesota. Avian Dis. 27, 77-85.
Hinshaw, V.S., Webster, R.G., Turner, B., 1980. The perpetu- ation of orthomyxoviruses and paramyxoviruses in Cana- dian waterfowl. Can. J. Microbiol. 26, 622-629.
Horimoto, T., Rivera, E., Pearson, J., Senne, D., Krauss, S., Kawaoka, Y., Wester, R.G., 1995a. Origin and molecular changes associated with emergence of a highly pathogenic H5N2 influenza virus in Mexico. Virology 213, 223-230.
Horimoto, T., Ito, T., Alexander, D.J., Kawaoka, Y., 1995b. Cleavability of hemagglutinin from an extremely virulent strain of avian influenza virus containing a unique cleavage site sequence. J. Vet. Med. Sci. 57, 927-930.
Inkster, M.D., Hinshaw, V.S., Schulze, I.T., 1993. The hemag- glutinins of duck and human H1 influenza viruses differ in sequence conservation and glycosylation. J. Virol. 67, 7436-7443.
Kawaoka, Y., Webster, R.G., 1988. Molecular mechanism of acquisition of virulence in influenza virus in nature. Mi- crob. Pathog. 5, 311-318.
Kawaoka, Y., Naeve, C.W., Webster, R.G., 1984. Is virulence of H5N2 influenza viruses in chickens associated with loss of carbohydrate from the hemagglutinin?. Virology 139, 303-316.
Lang, G., Narayan, O., Rouse, B.T., Ferguson, A.E., Connell, M.C., 1968. A new influenza virus A infection in turkeys. II. A highly pathogenic variant, A/turkey/Ontario/7732/66. Can. Vet. J. 9, 151-160.
Ludwig, S., Schultz, U., Mandler, J., Fitch, W.M., Scholtissek, C., 1991. Phylogenetic relationship of the nonstructural (NS) genes of influenza A viruses. Virology 183, 566-577.
Nettles, V.F., Wood, J.M., Webster, R.G., 1985. Wildlife surveillance associated with an outbreak of lethal H5N2 avian influenza in domestic poultry. Avian Dis. 29, 733- 741.
Perdue, M.L., Wainright, P., Brugh, M., 1990. Effects of chicken embryo age on time death following infection with avian influenza viruses: implications for distinguishing highly pathogenic isolates. Virus Res. 16, 137-152.
Rott, R., Orlich, M., Scholtissek, C., 1979. Correlation of pathogenicity and gene constellation of Influenza A viruses. III. Non-pathogenic recombinants derived from highly pathogenic parent strains. J. Gen. Virol. 44, 471- 477.
Saito, T., Horimoto, T., Kawaoka, Y., Senne, D.A., Webster, R.G., 1994. Emergence of a potentially pathogenic H5N2 influenza virus in chickens. Virology 201,277-284.
Senne, D.A., Pearson, J.E., Panigrahy, B., 1992. Live poultry markets: a missing link in the epidemiology of avian influenza. In: Proceedings 3rd International Symposium on Avian Influenza, Madison, WI, pp. 50-58.
Senne, D.A., Panigrahy, B., Kawaoka, Y., Pearson, J.E., Suss, J., Lipkind, M., Kida, H., Webster, R.G., 1996a. Survey of the hemagglutinin (HA) cleavage site: sequence of H5 and H7 avian influenza viruses: amino acid sequence at the HA
124 M. Garcia et al. / Virus Research 51 (1997) 115 124
cleavage site as a marker of pathogenicity potential. Avian Dis. 40, 425-437.
Senne, D.A., Rivera, E., Panigrahy, B., Fraire, M., Kawaoka, Y., Webster, R.G., 1996b. Characterization of avian influ- enza H5N2 isolates recovered from chickens in Mexico. In: Jensen, M.M., Ceniceros, M.A. (Eds.), Proceedings of the 45th Western Poultry Disease Conference, Cancun, Mex- ico. University of California, Davis, Conferences Services, pp. 5-8.
Slemons, R.D., Easterday, B.C., 1977. Type-A influenza viruses in the feces of migratory waterfowl. J. Am. Vet. Med. Assoc. 171, 947 948.
Slemons, R.D., Shieldcastle, M.C., Heyman, L.D., Bednarik, K.E., Senne, D.A., 199l. Type A influenza viruses in waterfowl in Ohio and implications for domestic turkeys. Avian Dis. 35, 165-173.
Stallknecht, D.E., Shane, S.M., Zwank, P.J., Senne, D.A., Kearny, M.T., 1990. Avian influenza viruses from migra- tory and resident ducks in coastal Louisiana. Avian Dis.
34, 398 405. Swofford, D. (1989). PAUP: Phylogenetic Analysis Using Par-
simony, Version 3. Illinois Natural History Survey, Cham- paign, IL.
Treanor, J.J., Snyder, M.H., London, W.T., Murphy, B.R., 1989. The B allele of the NS gene but not the A allele, attenuates a human influenza A virus for squirrel monkeys. Virology 171, 1-9.
Van Campen, H., Easterday, B.C., Hinshaw, V.S., 1989. Viru- lent avian influenza A viruses: their effect on avian lymphocytes and macrophages in vivo and in vitro. J. Gen. Virol. 70, 2887 2895.
Webster, R.G., Rott, R., 1987. Influenza virus A pathogenic- ity: the pivotal role of the hemagglutinin. Cell 50, 665-666.
Webster, R.G., Bean, W.J., Gorman, O.T., Chambers, T., Kawaoka, Y., 1992. Evolution and ecology of influenza A viruses. Microbiol. Rev. 56, 152-179.
Wood, G.W., McCauley, J.W., Bashiruddin, J.B., Alexander, D.J., 1993. Deduced amino acid sequences at the hemag- glutinin cleavage site of avian influenza A viruses of H5 and H7 subtypes. Arch. Virol. 130, 209-217.
Wood, G.W., Banks, J., McCauley, J.W., Alexander, D.J., 1994. Deduced amino acid sequences of the haemagglu- tinin of H5N1 avian influenza virus isolates from an out- break in turkeys in Norfolk, England. Arch. Virol. 134, 185 194.
